almost-khan/khanote/khanote-researcher-add

khanote.researcher.add

You are helping the user connect a new data source (API) to khanote.

Instructions

Walk the user through these steps conversationally — ask one question at a time.

Step 1: Describe the data source

Ask: "Describe the data source you want to connect. For example: 'I want to use Exa Search API to find research papers.'" Extract: API name, likely capabilities, and domain.

Step 2: Permanent or session-only?

Ask: "Save this researcher permanently for all future sessions, or just for this session?"

• Permanent → write to config.yaml
• Session only → write to .khanote/sessions/{slug}/researcher.yaml

Step 3: API configuration

Ask these three questions:

1. "What is the API endpoint URL?"
2. "How does this API authenticate? (API key in header, Bearer token, or no auth)"
3. "Do you have an API key? You can also use an environment variable like ${MY_API_KEY}."

Step 4: Capabilities

Ask: "What can this API do? Select all that apply: search, ingest, analyze, generate." Require at least one capability.

Step 5: Test the connection

Run khanote researcher add or call the API with a test request to verify the key and URL.

• Success: tell the user the connection works and show the response latency.
• Failure: show the error and ask: "Save anyway as unavailable? (yes/no)"

Step 6: Confirm and save

Show a summary of the researcher configuration and ask: "Save this researcher? (yes/no)"

• Yes: write to config.yaml (permanent) or session file.
• No: discard and exit cleanly.

Step 7: Offer to create a feed (permanent researchers only)

Ask: "Would you like to create a recurring feed using this researcher? (yes/no)"

• Yes: proceed with /khanote.feed.add with this researcher pre-selected.
• No: done.

Notes

• All configuration is done through this conversation — no CLI flags required.
• Your config is at {vault_path}/.khanote/config.yaml.